Many naturally occurring factors may contribute significantly to crop shortfalls such as epidemics of disease and pests and weather. In this talk, we discuss how modeling of historical epidemiological, weather, and soil data may be used to discern between regular seasonal variability and potential external shocks such as deliberate agro-terrorism. Early detection of these events requires the synthesis of all of these heterogeneous data streams to eliminate alternative explanations. We propose a Bayesian learning framework for learning the relationships between these factors, which we trained on a decade’s worth of historical observations. We demonstrate how this modeling methodology may then be used to identify earlier spatiotemporal anomalies in surveillance which are indicative of coordinated and preternatural dissemination of disease.
-Identification of novel data elements for monitoring food security
-Statistical & machine learning techniques to synthesize these disparate data streams for forecasting/prediction
-Discussion of metrics for detection for potential adverse events in food security
Andrew Hong, MITRE Corporation, Operations Research Analyst
Dr. Andrew E. Hong is an operations research analyst in the Modeling and Simulation department at the MITRE Corporation. He has supported several federal agencies such as DHS, IRS, and CMS in data-driven modeling for detecting fraud and waste.
Who can attend:
InfraGard members interested in the topic of detecting agro-terrorism with environmental analytics. Open only to InfraGard members and their guests.